historical ecology on sandoy, faroe islands ...€¦ · change on sandoy, faroe islands, especially...

Human Ecology, Vol. 33, No. 5, October 2005 ( C© 2005)DOI: 10.1007/s10745-005-7681-1

Historical Ecology on Sandoy, Faroe Islands:Palaeoenvironmental and ArchaeologicalPerspectives

Ian T. Lawson,1,12 Mike J. Church,2 Tom H. McGovern,3 Sımun V. Arge,4

James Woollet,5 Kevin J. Edwards,6 Freddy J. Gathorne-Hardy,7

Andrew J. Dugmore,8 Gordon Cook,9 Kerry-Anne Mairs,8

Amanda M. Thomson,10 and Gu run Sveinbjarnardottir11

We present palaeoenvironmental, geomorphological, archaeological, andplace-name data which allow a holistic assessment of the history of landscapechange on Sandoy, Faroe Islands, especially in terms of the changes that oc-curred in response to the colonization of the island by humans. In contrastto other situations in the North Atlantic region, there is considerable conti-nuity in the patterns and processes of landscape evolution across the initial

1Earth and Biosphere Institute, School of Geography, University of Leeds, Leeds LS2 9JT,United Kingdom.

2Department of Archaeology, Durham University, South Road, Durham DH1 3LE, UnitedKingdom.

3Anthropology Department, Hunter College CUNY, 695 Park Avenue, New York City10021, New York.

4Føroya Forminnissavn, Hoyvik, Postbox 1155, FO-110 Torshavn, Føroyar.5Departement d’Histoire, Pavillon Charles-DeKoninck, Universite Laval, Quebec, QC,Canada G1K7P4.

6Department of Geography and Northern Studies Centre, University of Aberdeen, St Mary’s,Elphinstone Road, Aberdeen AB24 3UF, United Kingdom.

7School of Conservation Sciences, Bournemouth University, Talbot Campus, Poole, DorsetBH12 5BB, United Kingdom.

8School of GeoSciences, University of Edinburgh, Drummond Street, Edinburgh EH8 9XP,United Kingdom.

9SUERC, Scottish Enterprise Technology Park, Rankine Avenue, East Kilbride, GlasgowG75 OQF, United Kingdom.

10School of Biological and Environmental Sciences, University of Stirling, Stirling FK9 4LA,Scotland, United Kingdom.

11Institute of Archaeology, University College London, 31-34 Gordon Square, London WC1H0PY, United Kingdom.

12To whom correspondence should be addressed; e-mail: [email protected].

651

0300-7839/05/1000-0651/0 C© 2005 Springer Science+Business Media, Inc.

652 Lawson et al.

settlement horizon. Many of the characteristic features of post-settlementNorth Atlantic landscapes—absence of trees, widespread blanket mires, highrates of soil erosion—were already in place when the first people arrived.Although human impact on Sandoy appears to have been light, conversely,the unusual environment forced major alterations of the subsistence econ-omy imported by the colonists. Settlement-era archaeological records suggestthat, from the start, patterns of resource use differed substantially from the re-gional norm, and these differences became amplified over time as the Faroeseeconomy created a locally sustainable cultural landscape.

KEY WORDS: Norse; human adaptation; palaeoecology; archaeology; landnam.

INTRODUCTION

Throughout the North Atlantic region, human colonization has causedgreat changes to the natural environment. In many areas, the arrival ofpeople, their agricultural systems, and their livestock, rapidly and radicallytransformed the flora, fauna, and soil system. This affected other inter-related aspects of the landscape, especially patterns of erosion and sedi-mentation, sometimes creating serious vulnerabilities to climate change anddrawing down the natural capital represented by pre-settlement soils, vege-tation, and wild animals. Perhaps the most dramatic example of human im-pact is that of Iceland, where colonization by the Norse in the ninth centuryA.D. led to deforestation and widespread soil erosion (e.g., Dugmore andBuckland, 1991; Hallsdottir, 1987; Olafsdottir, 2001; Simpson et al., 2001,2004). In Greenland, the long-term consequences of early human impactsmay have resulted in local settlement abandonment or even prejudiced thesurvival of the inhabitants by limiting adaptive options and reducing the re-silience of key natural and human systems (Barlow et al., 1997; McGovern,1997). In the case of the Faroes, the degree to which humans modified theenvironment is poorly known. Existing palaeoecological work is limited inits spatial coverage and mostly confined to pollen analysis, so that manyfactors—the extent and species composition of pre-settlement woodland,the impact of people on soils, even the date of the first colonization (sometime in the second half of the first millennium A.D.)—remain either disputed,or entirely unknown.

This paper attempts to establish how people and their activities fit intothe processes and patterns of landscape evolution on the Faroese islandof Sandoy. In the past, Sandoy has been largely overlooked by palaeoe-cologists, with the exception of a study on the early Holocene record ofGrothusvatn (Hannon et al., 2001). Here we look at the interrelated fac-tors of erosion, sedimentation, and vegetation change, principally throughinvestigations of the sedimentology, palynology, and palaeozoology of

Historical Ecology on Sandoy, Faroe Islands 653

two contrasting lakes. The palaeolimnological work is supported by de-tailed analyses of a number of peat, soil, and fluvial sequences, andby geomorphological mapping. Archaeobotanical and zooarchaeologicaldata from an excavation of Norse to early medieval middens at UndirJunkarinsfløtti in Sandur, together with place-name and historical evidence,provide data on the changing patterns of resource use by the settlers andtheir descendants. Comparisons with similar situations elsewhere in the re-gion suggest ways in which the imported Norse farming economy was mod-ified to suit the peculiarities of the Sandoy environment.

STUDY AREA

Sandoy, with an area of ca. 120 km2, lies close to the center ofthe Faroese archipelago (Fig. 1). Like all of the islands, the bedrock ofSandoy consists of stratified Tertiary basalts and associated tuffs, dipping

Fig. 1. Location map showing the island of Sandoy and sites discussed in the text. Grey shad-ing indicates topography at 50 m intervals; the highest peak on Sandoy is 479 m above sealevel. Inset shows the location of the Faroes.

654 Lawson et al.

gently eastwards (Rasmussen, 1982). The geology strongly determines thetopography, which is generally rugged, with high cliffs especially on thewestern coasts and west-facing hillsides. In the context of the Faroes, San-doy has unusually large tracts of low-lying, relatively flat land, includingthe bay and extensive valley at Sandur, and the large area of blanket peataround Lıtlavatn.

The present settlements are all coastal villages located near beachessuitable for landing boats. In the Sandur area, three primary holdings arethought to have been established by the Norse settlers, later subdividedinto the several smaller farms forming the basis of the modern village. Oneof these primary holdings has been the subject of archaeological excava-tion, producing radiocarbon dates which are currently the earliest settle-ment dates in the Faroes.

Each village on Sandoy is surrounded by an area of manured andperiodically cultivated infields, used for hay-making and the cultivationof barley (Hordeum sp.), and potatoes (Solanum tuberosum). Away fromthe infields, the vegetation is mostly a mix of grassland and, on peat, matgrass-crowberry-heather (Nardus stricta-Empetrum nigrum-Calluna vul-garis) moor (Fosaa, 2001). The rural economy is largely based on sheepfarming, although cattle were economically very important in the recentpast (Guttesen, 2001).

MATERIALS AND METHODS

Sediment sequences from two lakes on Sandoy were sampled using a10 cm diameter Russian-type corer from a boat, close to the center of eachlake. Grothusvatn (Figs. 1 and 2; Table I), situated in the coastal lowlandsnear the town of Sandur, has a relatively intensively managed catchmentat the present day, while Lıtlavatn (Fig. 3) lies at 60 m above sea level inwhat is today an outfield area used for rough grazing. Both lakes are small,shallow (2 m), and nutrient-poor, with sediments consisting of organic mudswith a variable component of clastic silicates and diatom frustules. A deeppeat sequence at Millum Vatna (Fig. 4) was sampled in a similar way, whileshallow blanket peats (e.g. Fig. 5) were sampled at nine locations within theLıtlavatn catchment using monolith tins.

Pollen extractions followed the standard technique outlined byBennett and Willis (2002), omitting steps 2, 4, 5, 6, and 9 in their schema,using Lycopodium tablets as a source of exotic markers, and silicone oil asa mounting medium. Samples were counted to a total of 300 (Lıtlavatn lakesequence and shallow peat sequences) or 500 (Grothusvatn, Millum Vatna)terrestrial pollen grains, excluding spores, aquatics, and alien taxa.


Fig

.2.S

elec

ted

taxa

polle

ndi

agra

mfo

rG

roth

usva

tn,S

ando

y.D

epth

sar

em

easu

red

belo

wth

ew

ater

surf

ace;

the

sedi

men

t-w

ater

inte

rfac

eis

at18

3cm

dept

h.T

heth

ick

dash

edlin

ehe

re,a

ndin

Fig

s.3

and

4,in

dica

tes

appr

oxim

atel

yth

ele

vel

abov

ew

hich

the

data

indi

cate

anth

ropo

geni

cac

tivi

ty.B

lack

dots

indi

cate

the

pres

ence

ofpo

llen

atfr

eque

ncie

sbe

low

0.5%

.

656 Lawson et al.

Table I. Radiocarbon Dates From the Grothusvatn, Lıtlavatn, Millum Vatna, and Lıt-6Sequences

14C yr B.P.Code Site and sample Sample type δ13C (1σ error)

SUERC-1829 Grothusvatn 248–249 cm Gyttja −27.2 2955 ± 40SUERC-1830 Grothusvatn 271–272 cm Gyttja −27.3 3050 ± 40SUERC-1821 Lıtlavatn 36–37 cm Gyttja −27.5 2980 ± 45SUERC-1826 Lıtlavatn 67–68 cm Gyttja −26.8 2125 ± 40SUERC-1827 Lıtlavatn 110–111 cm Gyttja −25.0 4200 ± 35SUERC-1828 Lıtlavatn 210–211 cm Gyttja −22.1 7430 ± 45SUERC-1831 Millum Vatna 45–46 cm Peat (humic −28.2 1375 ± 35

fraction)SUERC-1832 Millum Vatna 113–114 cm Peat (humic −27.8 1595 ± 35

fraction)SUERC-167 Lıt-6 Peat (humic −27.4 4425 ± 25

fraction)

Loss-on-ignition (Dean, 1974) and, in some cases, magnetic suscep-tibility (Dearing, 1999) analyses were also performed, the latter using aBartington Instruments MS-2 meter and MS2b sensor.

Fossil remains of the head capsules of chironomid larvae were analyzedfrom the Grothusvatn and Lıtlavatn sequences following the methodologyoutlined by Lang et al. (2003). Total phosphorus reconstruction was car-ried out following Brooks et al. (2001), using a 44-lake training set from theEnglish Midlands and Wales (UK).

The archaeological data presented here come primarily from a site onSandoy, at Undir Junkarinsfløtti in Sandur (Fig. 1). The site consists of aseries of middens, wall lines, and buried structures, dated to the ninth–earlythirteenth centuries A.D., and partly exposed by coastal erosion of the sandysoil of the modern infield. The middens were first identified in 2000 afterslumping caused by a prolonged dry period and an initial trial-trenchingexercise recorded a series of bone- and ash-rich middens over 2 m thick.Two radiocarbon determinations (Table II) from the two lowest middendeposits produced dates in the ninth–tenth century A.D. The early datingof the lowermost midden material was reinforced by the discovery of abronze brooch of tenth century A.D. date in the same layer (Arge, 2001). Alarger-scale sampling exercise was undertaken in 2003 to enlarge the smallsondage excavated in 2000 with a view to extracting zooarchaeological andarchaeobotanical remains. The details of the investigation are presented byChurch et al. (2005) with summary findings outlined here. All of the archae-ological deposits were dry-sieved at 4 mm for the extraction of zooarchae-ological remains and artefacts. Bulk samples were taken from every con-text (total sampling: Jones, 1991) from which archaeobotanical and smallerzooarchaeological remains were extracted.


Fig

.3.S

elec

ted

taxa

polle

ndi

agra

mfo

rL

ıtla

vatn

(lak

ese

quen

ce).

Dep

ths

are

mea

sure

dbe

low

the

sedi

men

tsur

face

.

658 Lawson et al.

Fig

.4.S

elec

ted

taxa

polle

ndi

agra

mfo

rM

illum

Vat

na.D

epth

sar

em

easu

red

belo

wth

epe

atsu

rfac

e.


Fig. 5. Selected taxa pollen diagram for Lıt-6, a typical short peat/soil sequence fromthe area of blanket peat around Lıtlavatn. Depths are measured below the peat surface.The loss-on-ignition curve shows a rapid increase in organic content between 45 and41 cm as mineral soil gives way to peat, mirroring a shift in the pollen from herbaceouscommunities to ericaceous taxa. The radiocarbon date, from the lowermost level witha high organic content, is a conservative estimate of the age of peat initiation at thissite. Similar measurements on eight other sequences from the Lıtlavatn area consistentlyyielded pre-landnam ages for peat initiation.

All radiocarbon calibrations were performed using the computer pro-gram OxCal (Bronk Ramsey, 2003), with the INTCAL98 calibration dataset (Stuiver et al., 1998).

VEGETATION

The paleoenvironmental analysis of the lake and peat sequences out-lined above allows an assessment of the environment of the Faroes as itwould have appeared to the first settlers.

660 Lawson et al.

Table II. Radiocarbon Dates From Undir Junkarinsfløtti, Sandur

Code Context Sample type δ13C 14C yr B.P. (1σ error)

SUERC-3422 3 Barley grain −24.6 925 ± 40SUERC-3417 16 Barley grain −25.9 900 ± 35SUERC-3418 16 Barley grain −26.4 925 ± 40AAR-6927 19 Sheep bone −19.8 950 ± 35SUERC-3423 22 Cow bone −20.9 990 ± 35SUERC-3424 22 Pig bone −21.2 1035 ± 35SUERC-3400 23 Barley grain −23.9 1000 ± 40SUERC-3401 23 Barley grain −26.8 980 ± 40SUERC-3402 23 Barley grain −26.3 940 ± 45SUERC-3403 23 Barley grain −24.0 995 ± 35SUERC-3410 23 Pig bone −21.4 965 ± 40SUERC-3411 23 Pig bone −21.0 1075 ± 40SUERC-3415 23 Pig bone −21.4 935 ± 40SUERC-3416 23 Pig bone −21.6 1005 ± 35SUERC-3425 23 Cow bone −21.0 980 ± 40SUERC-3426 23 Pig bone −22.7 1095 ± 40AAR-6929 23 Cow bone −19.9 1115 ± 35AAR-6928 24 Sheep bone −20.4 1190 ± 40

The Pre-settlement Vegetation of Sandoy

The age of the blanket peat on Sandoy has been determined byradiocarbon dating of the oldest peats in the short peat/soil sequences,which typically yielded dates of 3000–5000 14C B.P. (approximately 1000–4000 cal B.C.; Fig. 5). The mid-Holocene expansion of peatland commu-nities dominated by Empetrum nigrum and Calluna vulgaris is also regis-tered in the Grothusvatn pollen record, and less clearly in the Lıtlavatnsequence where periodic reworking obscures the pollen data. Thus peataccumulation began long before the first arrival of people, presumablythrough the well-established process of progressive leaching of nutrientsand acidification as the soils matured through the Holocene (e.g., Crawford,2000; Charman, 2002). This result stands in contrast to many situations inthe North Atlantic region where human agency has been implicated in peatinitiation (e.g., Bennett et al., 1997; Bunting, 1996; Charman, 1992; Moore,1975, 1993; Solem, 1989).

Today the Faroes are treeless, apart from a few plantations. Althoughtree birch (Betula pubescens) macrofossils are occasionally found in pre-settlement peats on other islands (Malmros, 1994), previous pollen datahave shown no indication that woodland was ever abundant. Tree birch per-centages in pre-landnam samples are typically small in the various pollenspectra from Sandoy (usually 0.7–3.9% total land pollen). These valuesalone would be consistent with either a very low density of local woodland,


or long-distance dispersal from Britain or the European mainland (cf.Randall et al., 1986; Brayshay et al., 2000; Tyldesley, 1973). However, a sin-gle sample from a short peat/soil profile in the Lıtlavatn catchment, Lıt-7,contains 16.3% B. pubescens pollen. This very high value suggests thatbirch is likely to have been present at least intermittently on Sandoy dur-ing the Holocene, while the absence of large amounts of its pollen in theother sequences suggests that it was never widespread or abundant, per-haps occurring only in localities protected from salt spray. No other treetaxon has yet been found in such abundance in the pollen record as to sug-gest local presence.

Overall, immediately pre-settlement pollen assemblages suggest a rea-sonably diverse landscape, with a mosaic of grasses and sedges, juniperscrub, and ericaceous heath and mire communities. There is also evidencethat tall herbs such as Angelica sylvestris (wild angelica) and Filipendula ul-maria (meadowsweet), typical of ungrazed situations (e.g., Bennett et al.,1992), were relatively abundant.

Vegetation Change After the Settlement

The clearest record of post-settlement vegetation change on Sandoyis the pollen record from Grothusvatn (Fig. 2). Identification of the set-tlement horizon in the record is complicated by the difficulty of datingthe sediments. Radiocarbon ages of the total carbon fraction from 1 cm3

of gyttja from both Grothusvatn and Lıtlavatn appear (on the basis ofbiostratigraphic correlation with other dated sequences from the Faroes;Hannon et al., 1998, 2001; Hannon and Bradshaw, 2000; Johansen, 1975,1982, 1985) to be affected by redeposition of older organic matter from thelake catchments, giving dates that are too old. Similarly, redeposition oftephra hinders the application of tephrochronology as an alternative dat-ing tool. Biostratigraphic correlation with the more reliably dated MillumVatna sequence, and other well-dated sites from across the Faroes, suggeststhat the base of the Grothusvatn sequence is older than ca. 5000 B.P., and al-lows the tentative identification of the settlement horizon at ca. 260–270 cmdepth in the sequence. Here, the presence of pronounced and broadly syn-chronous changes in several taxa following a long period of essentiallystable assemblages suggests a major perturbation, while the presence ofcharcoal above this level is a strong indication that the perturbation is an-thropogenic.

Above 260–270 cm, significant changes in the pollen assemblages arerecorded, including a gradual decline in Juniperus communis (juniper),similar to that found at Tjørnuvık (Hannon et al., 1998; Hannon and Brad-shaw, 2000); a steady rise in Plantago, especially P. lanceolata (ribwort

662 Lawson et al.

plantain), similar to those found at Hovi and Tjørnuvık (Johansen, 1971,1975, 1985); and a gradual increase in the abundance of Poaceae (grasses)with a corresponding decline in Calluna vulgaris (heather), especially above215 cm. These changes suggest the progressive removal of Juniperus bygrazing or collection for its various uses (Malmros, 1990, 1994; Small, 1992);a spread of anthropochorous weeds like Plantago spp. responding to soildisturbance; and the replacement of ericaceous (heath) taxa by grassesthrough some combination of the favouring of plants such as Nardus stricta(mat-grass) by grazing on the outfields (Fosaa, 2001; Johansen, 1985; Nolanet al., 1995; Stevenson and Thompson, 1993;), and the displacement ofheath communities in the infields by cultivation. Very occasional finds ofcereal-type grains suggest cultivation of crops; all grains found so far areattributable to Hordeum-type.

The changes in the pollen assemblages at Grothusvatn are not sud-den, as, for example, at the settlement horizons at Mosfell or randarholt inIceland (Hallsdottir, 1987), or at Dallican Water in the Shetlands (Bennettet al., 1992). A major difference in the case of the Faroes is that, inthe absence of significant tracts of woodland, deforestation with all itsknock-on effects on other vegetation communities is simply an impossi-bility. However, pollen data from infield sites in the Faroes, for exampleHovi and Tjørnuvık (Edwards et al., 2005; Hannon et al., 1998; Hannonand Bradshaw, 2000, Johansen 1971, 1975, 1985), also show a more suddenchange than at Grothusvatn. A possible explanation is that Grothusvatn,being a shallow lake, is likely to experience mixing and redeposition of itssediments through wind action, effectively blurring the sediment sequence.Counting against this argument is the observation that pollen data from thepeat sequence at Millum Vatna likewise indicate a gradual change in, forexample, Juniperus across the settlement horizon. It may be the case thatimpacts were sudden where land management was intensive, as at the in-field sites at Hovi and Tjørnuvık, but more progressive in the outfield areas,which are more strongly represented in the Grothusvatn and Millum Vatnasequences.

At Lıtlavatn, redeposition of pollen obscures the pattern of vegetationchange, although P. lanceolata becomes more common towards the top ofthe sequence. At Millum Vatna a similar, gradual increase in P. lanceolata,associated with small amounts of charcoal, also occurs towards the top ofthe sequence. At this site radiocarbon dating supports the identification ofthe settlement horizon as approximately coincident with these changes. It isnot clear whether other changes in the Millum Vatna sequence, for exam-ple the increasing ratio of Poaceae to Calluna, are natural or anthropogenic(cf. Stevenson and Thompson, 1993), given that peat sequences often re-cruit much of their pollen from a very small area (Bunting, 2003) and hence


are capable of recording small-scale, natural shifts in the vegetation mosaic.Juniperus pollen appears even in the next-to-uppermost sample analyzed,but it was apparently not abundant at this site even before landnam.

In short, the pollen records from Sandoy suggest that anthro-pogenic changes to the vegetation were both subtle and gradual. NeitherGrothusvatn nor Millum Vatna shows evidence for abrupt change. Fur-thermore, the data suggest that the changes to the vegetation composi-tion were fairly insubstantial, at least on a landscape scale, with no greatstructural change such as the removal of dense woodland, or the initiationof peat over large areas, as seen in other areas of the North Atlantic. Atmost, soil disturbance encouraged an expansion of ruderal herbs, grassesexpanded slightly in response to grazing, and juniper bushes gradually dis-appeared from the landscape, along with the rare stands of tree birch. Animportant qualification to this conclusion is that the pollen records gener-ated so far are either integrative over large areas, or situated in the leastheavily exploited, and possibly least fragile areas of the landscape. Humanimpact was presumably, in the past as now, concentrated on relatively smallareas. Comparison with the results of Johansen (1982, 1985) from Hovi, andof Johansen (1985) and Hannon and Bradshaw (2000; also Hannon et al.,1998, 2001) at Tjørnuvık, and palaeoentomological work in Tjørnuvık andToftanes (Buckland et al. 1998, Vickers et al., 2005), suggests that change inat least some infield areas and farmyards was likely to have been more sub-stantial and more abrupt. As yet, little is known about vegetation change atmoderate to high altitudes in the Faroes (although cf. Edwards et al., 2005),where the lower productivity of the vegetation might lead to a greater sen-sitivity to human disturbance.

EROSION, SEDIMENTATION, AND AQUATIC ENVIRONMENTS

A typical response of landscapes to human settlement is an increase inerosion (Edwards and Whittington, 2001). This often occurs via destructionof the vegetation that binds the topsoil together, whether by cultivation, de-forestation, overgrazing, or trampling. The history of erosion on Sandoy isbeing reconstructed through a study of limnic and fluvial sediments, alluvialfans, and soils and peats.

Palaeolimnology

Loss-on-ignition and magnetic susceptibility data for Lıtlavatn andGrothusvatn provide continuous records of erosion in their catchments,

664 Lawson et al.

revealing very different sedimentation histories for the two lakes. The mag-netic susceptibility record for Grothusvatn (Fig. 2), which can be viewed as aproxy for soil erosion in the catchment (Dearing, 1999), indicates little min-eral input and no significant long-term trends. At Lıtlavatn (Fig. 3), by con-trast, the magnetic susceptibility data show large pulses of mineral sedimentinwash, including one at the base of the sequence, one at 70–115 cm, andanother at 0–35 cm. (Loss-on-ignition data for the two lakes generally ac-cord with the magnetic susceptibility data, although occasionally abundantdiatom frustules account for deviations.) A partial explanation of the differ-ences between the two sedimentary records is that, although the catchmentsof the two lakes are approximately the same size (ca. 470 and 460 ha respec-tively; Fig. 6), the surface area of Lıtlavatn is about half that of Grothusvatn,so a volumetrically equal flux of sediment through both catchments will bemore evident in the deposits of the smaller lake. In Grothusvatn, the trans-port of bedrock- or soil-derived mineral material into the lake is furtherlimited by the presence of a small basin just upstream of the lake on itsmain tributary stream, which acts as a sediment trap, and by a fen at thenorthern end of the lake which filters out some of the remaining sediment.Nonetheless, the conclusion to be drawn from the Lıtlavatn record is thatsubstantial erosion had occurred well before the arrival of people; in fact,according to the biostratigraphy, actually before the mid-Holocene expan-sion of Calluna. Erosion rates increase again towards the top of the se-quence, probably in the post-settlement period, perhaps relating to LittleIce Age climatic change (see below). The data thus show that erosion on asignificant scale, at least in the uplands, was part of the natural landscape ofSandoy well before the first settlement, although human activity may haverevitalized the process.

Changing rates of erosion may have had substantial impacts on ter-restrial flora and fauna, providing niches for pioneer taxa, although thisremains to be demonstrated (perhaps the rise of Plantago lanceolata at2500 B.P. is an indication of increased natural disturbance, not the immi-gration of this taxon; cf. Johansen, 1989). There is evidence that changeswere also occurring in the lake ecosystems. At Grothusvatn, total phos-phorus estimates based on changes in chironomid assemblages in the up-per part of the sequence indicate a substantial rise in the nutrient statusof the lake following landnam. This conclusion is supported by an increas-ing abundance of the alga Pediastrum (Fig. 2), closely tracking the totalphosphorous reconstruction. Together, these data suggest that an increasein lake productivity accompanied the settlement, perhaps due to mobiliza-tion of nutrients by grazing and manuring, or to soil erosion.

At Lıtlavatn, the change in sedimentation regime in the mid-Holocene is accompanied by dramatic declines in productivity by diatoms,


Fig. 6. Frequency histograms to compare the topographic characteristics of the hy-drological catchments of Grothusvatn and Lıtlavatn. (a) Area of land at differentaltitudes (height classes) in the two catchments. The Lıtlavatn catchment is, on av-erage, considerably higher than that of Grothusvatn. (b) Area of land of differentslope angles in the two catchments. The Lıtlavatn catchment includes more steeply-sloping land, on which soils are more readily eroded. These differences help to ac-count for the fact that the two lakes record markedly different sedimentologicalhistories. The histograms were generated using data from quantitative analysis ofdigitized topographic data.

Pediastrum, and chironomids. Contrary to the increasing productivity atGrothusvatn, sustained high levels of clastic input seem to have inhibitedthe productivity of this lake, a situation that could only be exacerbated iflevels of erosion increased still further following settlement.

666 Lawson et al.

Geomorphology

A regionally representative stratigraphic sequence has been identifiedon Sandoy from observation and sampling of over 50 sediment profilesacross the landscape. The sequence at this stage is based on field observa-tion, with laboratory and dating analysis still ongoing, but broad trends canbe identified within the context of a Holocene chronology. The sequencebegins with peat accumulation overlying a basal early Holocene minero-genic soil and/or glacial diamicton and/or bedrock. The peat cover is exten-sive and generally absent only from heavily eroded and high altitude areas.In lowland areas and on slopes, the peat is overlain by minerogenic sedi-ment formed by an influx of sands and gravels, represented in the profileseither as a single unit or as a series of laminations of coarse sand and gravelinterdigitated with finer organic silts. This unit is an indication of exten-sive geomorphic change and landscape instability, represented by increasedslope erosion and re-deposition by wind and water. This sand/gravel unit isin turn overlain by an organic silty soil, indicating a change in slope pro-cesses with less energetic remobilisation of sediment. This stratigraphic se-quence is representative of many of the recorded profiles on both Su uroyand Sandoy.

Based on chronological information from past palaeoecological re-search from the Faroes (cf. Johansen, 1985; Hannon et al., 2001; Humlumand Christiansen, 1998; Wastegard, 2002), the sequence indicates that inthe mid-Holocene, the Faroe Islands were more geomorphologically stablethan today, with a wet climate facilitating extensive peat formation. Thepeat developed to cover much of lowland Sandoy. In the later Holocene(2000–4000 cal. B.P.) these lower slopes destabilized; the identification of themechanisms and processes of this destabilization is still underway, but in theabsence of people and large herbivores this shift is likely to be a result ofclimate change that crossed a geomorphic threshold, although the changingpatterns of large colonies of nesting birds, particularly puffins, may also berelevant. This environmental change disrupted parts of the extensive low-land peat cover, leading to the creation of a more varied “mosaic” of vege-tation and landscapes. At the same time, peat continued to develop acrossthe island (cf. Humlum and Christiansen, 1998). The first settlers thus en-countered a markedly dynamic landscape in which vegetation disturbance,changing patterns of runoff, gullying, soil and peat erosion, and alluviationwere all well-established processes.

Since the settlement a top unit of organic silt has developed acrossmuch of Sandoy, indicating a change in geomorphic activity, a phase of sedi-ment mobilization perhaps related to the combined effects of Little Ice Ageclimate change and human impact. Changes in atmospheric circulation in


the early fifteenth century led to increased storminess in the North Atlantic(Dugmore et al., in press; Meeker and Mayewski, 2002), and more frequentand/or more intense rainfall, combined with periods of lower temperatureand human impact on vegetation and soils, could have enhanced processesof erosion and transport, leading to the deposition of this silt layer.

ARCHAEOLOGY

A composite section of most of the archaeological deposits at thekey site of Undir Junkarinsfløtti is presented in Fig. 7, running from thesterile sand overburden (Context 2) down to the lowest midden deposits(Contexts 23 and 24), overlying the pre-settlement soil interface with theglacially-derived subsoil (Context 25). The deposits are characterized by

Fig. 7. Composite section drawing from Undir Junkarinsfløtti, Sandur.

668 Lawson et al.

concentrations of burnt and unburnt animal bone, limpet shells, and flecksof wood charcoal in a matrix composed of shell sand, peat ash, and fist-sized fire-cracked stones. The deposits excavated in 2003 appear to repre-sent refuse dumped from structures uphill (Woollett et al., 2004). Furtherexcavations just uphill from the eroding edge in 2004 have revealed a sub-stantial Late Norse stone structure partially filled with midden just abovethe level of the 2003 midden deposits.

Table II presents the radiocarbon determinations obtained from thesite and Fig. 8 presents the calibrated dates. The site has been separatedinto three phases based on the archaeological stratigraphy, radiocarbondates and artefacts. UJF1 represents the earliest deposits (Contexts 21–28)dated to ninth–twelfth centuries A.D., UJF2 includes Contexts 15–20 dat-ing to eleventh–twelfth centuries A.D., and UJF3 includes Contexts 3–14and dates to the eleventh–early thirteenth centuries A.D. The formation ofthe overlying post-thirteenth century A.D. sterile sand cover (Context 2) isconsistent with climate changes that mark the onset of the Little Ice Age inthe early thirteenth century (Dugmore et al., in press).

Plant Resource Use

Archaeobotanical materials from the three phases at UndirJunkarinsfløtti are scarce and consist of carbonized plant macrofossilsof cereal grains, a little cereal chaff, various plant parts from wild species,a few small pieces of charcoal, burnt peat and some turf (Table III).The presence of abundant charred peat and turf fragments and lowconcentrations of charcoal indicate that peat and turf were the primaryfuel sources. Some of the wild species, such as Calluna vulgaris and smallculm bases/rhizomes, were probably introduced by the use of peat and turfas fuel (Church, 2002; Dickson, 1998). Peat procurement was thus a keycomponent in the Faroese Norse economy, and its extraction would havehad a visible impact on the landscape.

The cereal remains are dominated by six-row hulled barley (Hordeumvulgare var. vulgare), the staple cereal of the Norse period in the NorthAtlantic, with a few grains of oat (Avena sp.). Oat grains cannot be identi-fied to species without their floret bases, which did not survive at the site,and so it impossible to determine whether the oats were introduced as aweed of the barley, or cultivated. The barley would have been grown in theinfield area, although there is a possibility of grain being imported to theislands. The ubiquitous presence of chickweed (Stellaria media) in all threephases indicates relatively nitrogenous soil conditions (Sobey, 1981). If thechickweed was a weed of the barley crop (though the plant is associated


Fig. 8. Calibrated radiocarbon dates from Undir Junkarinsfløtti. The phasing of the site isbased on stratigraphic observation and artifactual dating as well as the radiocarbon dating(Church et al., 2005).

with other plant communities as well as cultivated land; Fosaa, 2000, 2001),this may represent field rotation between pastoral and arable agricultureon a seasonal or spatial basis, or the deliberate incorporation of dung intothe soil as a fertilizer and stabilizer. The remaining wild species represent amixture of taxa from cultivated ground, wet pasture, and moorland. Theseplants, from mutually exclusive ecological niches (Grime et al., 1988), wouldhave been introduced to the domestic hearths by the burning of peat and

670 Lawson et al.

Tab

leII

I.Su

mm

ary

Arc

haeo

bota

nica

lRem

ains

Fro

mU

ndir

Junk

arin

sfløt

ti,S

andu

r

Pha

seU

JF1a

Pha

seU

JF2b

Pha

seU

JF3c

Tot

alsi

ted

Cha

rcoa

lB

etul

asp

.tim

ber

frag

men

t7F

(1.0

2)7F

(1.0

2)C

allu

navu

lgar

is(L

.)ro

undw

ood

(2–4

mm

)p

pp

Cal

luna

vulg

aris

(L.)

roun

dwoo

d(4

+m

m)

1F(0

.04)

1F(0

.04)

Con

ifer

aein

det.

tim

ber

frag

men

t2F

(0.0

3)2F

(0.0

3)Ju

nipe

ris

sp.r

ound

woo

d1F

(0.0

3)1F

(0.0

3)L

arix

sp.t

imbe

rfr

agm

ent

6F(0

.15)

2F(0

.04)

2F(0

.05)

10F

(0.2

4)P

icea

sp.t

imbe

rfr

agm

ent

1F(0

.02)

1F(0

.02)

Pin

ussp

.tim

ber

frag

men

t1F

(0.0

3)1F

(0.0

3)Q

uerc

ussp

.tim

ber

frag

men

t4F

(0.0

9)4F

(0.0

9)B

urnt

peat

/tur

ffra

gmen

ts15

4.5

g53

.2g

128.

6g

336.

2g

Gra

inH

orde

umsp

.(C

)B

arle

ygr

ain

1730

754

H.h

ulle

d(C

)H

ulle

dba

rley

grai

n13

123

28H

.hul

led

sym

met

ric

(C)

Hul

led

barl

eyst

raig

htgr

ain

31

37

H.h

ulle

das

ymm

etri

c(C

)H

ulle

dba

rley

twis

ted

grai

n7

512

Ave

nasp

.(C

)O

atgr

ain

11

2C

erea

lind

eter

min

ate

(C)

Cer

ealg

rain

3251

891

Tot

algr

ain

7310

021

194

Cha

ffH

orde

umsp

.(R

I)B

arle

yra

chis

12

3H

.vul

gare

L.(

RI)

Six

row

barl

eyra

chis

11

Cer

eal/m

onoc

otyl

edon

(>2

mm

.)(C

N)

Cer

eal/m

onoc

otyl

edon

culm

node

11

Cer

eal/m

onoc

otyl

edon

(>2

mm

.)(C

B)

Cer

eal/m

onoc

otyl

edon

culm

base

11

Tot

alch

aff

21

36

Wild

plan

tsB

rass

ica/

Sina

pis

spp.

(S)

Cab

bage

/Mus

tard

11

Cal

luna

vulg

aris

(L.)

Hul

l.(L

F)

Lin

ghe

athe

r2F

2FC

arex

spp.

(tri

gono

us)

(N)

Sedg

e3

14

Dan

thon

iade

cum

bens

L.(

C)

Hea

th-g

rass

11

Mon

tiafo

ntan

aL

.(S)

Blin

ks1

23

Historical Ecology on Sandoy, Faroe Islands 671T

able

III.

Con

tinu

ed

Pha

seU

JF1a

Pha

seU

JF2b

Pha

seU

JF3c

Tot

alsi

ted

Poa

ceae

undi

ff.(

med

ium

)(C

)G

rass

11

2P

oace

aeun

diff

.(sm

all)

(C)

Gra

ss1

21

4P

olyg

onum

spp.

(N)

Kno

tgra

ss1

1R

anun

culu

sre

pens

L.(

A)

Cre

epin

gbu

tter

cup

66

Ran

uncu

lus

spp.

(A)

But

terc

up1

12

Rum

exac

etos

aL

.(N

)C

omm

onso

rrel

11

2R

umex

cris

pus/

obtu

sifo

lius

L.(

N)

Cur

led

dock

21

58

Rum

exsp

p.(N

)D

ock

11

2Sp

ergu

laar

vens

isL

.(S)

Cor

n-sp

urre

y1

910

Stel

lari

am

edia

(L.)

Vill

ars

(S)

Com

mon

chic

kwee

d6

107

23V

iola

sp.(

S)V

iole

t1

1C

erea

l/mon

ocot

yled

on(<

2m

m)

(CN

)C

erea

l/mon

ocot

yled

oncu

lmno

de3

32

8C

erea

l/mon

ocot

yled

on(<

2m

m)

(CB

)C

erea

l/mon

ocot

yled

oncu

lmba

se4

52

11In

dete

rmin

ate

(>2

mm

)(R

)In

dete

rmin

ate

rhiz

ome

12

3In

dete

rmin

ate

(<2

mm

)(R

)In

dete

rmin

ate

rhiz

ome

11

46

Inde

term

inat

e(t

rigo

nous

)(S

/F)

Inde

term

inat

etr

igon

ous

seed

/fru

it1

12

Inde

term

inat

epe

rica

rpfr

agm

ent(

P)

Inde

term

inat

epe

rica

rpfr

agm

ent

11

Inde

term

inat

ese

ed/f

ruit

(S/F

)In

dete

rmin

ate

seed

/fru

it8

511

24M

oss

frag

men

ts(c

arbo

nise

d)(L

F)

Mos

sle

affr

agm

ent(

carb

onis

ed)

1F1F

Tot

alw

ild38

3057

125

Tot

alqu

anti

fiabl

eco

mpo

nent

s11

313

181

325

Gra

in/li

tre

1.7

1.6

0.4

1.2

Qua

ntifi

able

com

pone

nt/li

tre

2.6

2.1

1.4

2.0

Not

e.K

eyto

plan

tpar

ts:C

harc

oal:

F=

frag

men

t;p

=pr

esen

t;fig

ures

inbr

acke

tsgi

veto

talm

ass

ingr

ams.

Gra

in:(

C)=

cary

opsi

s.C

haff

:(C

B)=

culm

base

(>2

mm

diam

eter

);(C

N)

=cu

lmno

de(>

2m

mdi

amet

er);

(RI)

=ra

chis

inte

rnod

e.W

ildpl

ants

:(A

)=

ache

ne;

(C)

=ca

ryop

sis;

(CB

)=

culm

base

(<2

mm

diam

eter

);(C

N)

=cu

lmno

de(<

2m

mdi

amet

er);

(F)

=fr

uit;

(LF

)=

leaf

frag

men

t;(N

)=

nutl

et;(

P)

=pe

rica

rp;

(R)=

rhiz

ome;

(S)=

seed

.Nom

encl

atur

efo

llow

sSt

ace

(199

7).

aN

umbe

rof

sam

ples

inph

ase

=5;

Tot

alvo

lum

eof

sam

ples

=44

L.

bN

umbe

rof

sam

ples

inph

ase

=6;

Tot

alvo

lum

eof

sam

ples

=63

L.

c Num

ber

ofsa

mpl

esin

phas

e=

7;T

otal

volu

me

ofsa

mpl

es=

58.5

L.

dN

umbe

rof

sam

ples

inph

ase

=18

;Tot

alvo

lum

eof

sam

ples

=16

5.5

L.

672 Lawson et al.

turf as fuel. This mixing is a common feature of archaeobotanical assem-blages in the North Atlantic (Church and Peters, 2004) and negates detailedanalysis of crop weed ecologies.

The few charcoal remains included heather and juniper roundwood,presumably representing native growth (Fosaa, 2000); the few birch tim-ber fragments are probably native; and various coniferous timber speciesof larch (Larix sp.), pine (Pinus sp.), and spruce (Picea sp.) would have ar-rived on the island as driftwood. Driftwood was commonly used by Norsesocieties where native trees were scarce (Dickson, 1992; Kristjanson, 1980;Malmros, 1994), and its exploitation was regulated by legislation in earlymedieval Iceland (Dennis et al., 2000; pp. 321–343).

Animal Resource Use

The substantial, well-preserved 2003 archaeofauna from UndirJunkarinsfløtti (Table IV) provides some of the first zooarchaeological evi-dence for early economic strategies in the Faroes.

Figure 9 shows the distribution of major taxa in the three phases ofthe archaeofauna. In each case, bones of domestic and marine mammalsmake up a small portion of the collection (ca. 2–5%) compared to the largeamount of bird, fish, and shellfish remains. Birds (mainly puffin, Fraterculaarctica) come to outnumber fish bones in the upper layers, while shellfish(mainly limpets, Patella vulgaris) also increase in the upper layers.

Figure 10 shows the three major contexts from Junkarinsfløtti along-side contemporary archaeofauna from Iceland and Greenland. In Iceland,many landnam-era collections are also dominated by wild species, partic-ularly birds, but even at first settlement (Herjolfsdalur: Amorosi et al.,1996; Sveigakot: McGovern et al., 2001; Tjarnargata 4: Perdikaris et al.,

Table IV. Summary Archaeofauna From Undir Junkarinsfløtti, Sandur

UJF 1 UJF2 UJF 3 Total

Domestic mammals 76 118 241 435Whales 1 2 0 3Seals 1 6 7 14Birds 1,068 1,167 2,148 4,383Fish 2,400 573 1,157 4,130Shellfish 183 268 1,029 1,480NISP 3,729 2,134 4,582 10,445Medium (dog-pig-sheep sized) terrestrial 98 176 289 563

mammalLarge (cow-horse sized) terrestrial mammal 16 3 11 30Unidentified fragments 980 1128 2151 4,259TNF 4,823 3,441 7,033 15,297

Note. NISP: number of identified specimens; TNF: total number of fragments.


Fig. 9. Distribution of major taxa in the three phases of the 2003 Undir Junkarinsfløtti ar-chaeofauna. The ‘mammals’ category mostly consists of domestic mammal taxa, but includesa small contribution from whales and seals which is not visible when plotted at this scale.

2002) domestic mammal bone percentages are normally 15% or above.Wild species quickly decline in importance after the settlement, with do-mestic mammals making up 40–70% of collections (Sveigakot, Hofsta ir,Hrısheimar, Selhagi: McGovern et al., 2001; Svalbar : Amorosi, 1992) bythe tenth–eleventh century A.D. While the Greenlandic Norse colonists(sites W48: McGovern et al., 1983; W51: McGovern et al., 1996; E17a:McGovern et al., 1993; and GUS [Garden Under Sandet], Enghoff, 2003)made considerable use of seals and caribou throughout their occupation ofGreenland, their archaeofauna still show 15–40% domesticates (McGov-ern, 1985; Outram, 1999, 2003). The emphasis on wild species in general,and birds in particular, from the ninth to twelfth/thirteenth centuries atUndir Junkarinsfløtti is therefore unusual in the regional context. It hasbeen argued that sea bird colonies in southern Iceland represented a storeof natural capital that was quickly drawn down by the settlers, helping tosustain them until their imported stock could multiply (Vesteinsson et al.,2002). At Undir Junkarinsfløtti, by contrast, exploitation of sea birds wassustained over the long term.

Domestic stock-raising practices apparently changed only slightly onSandoy from the Norse to Medieval periods. Figure 11 presents thechanging proportions of domestic mammal bones from the 2003 UndirJunkarinsfløtti collection. The relative proportion of cattle decreases be-tween the earliest and subsequent phases, a pattern widely observed in mostNorth Atlantic landnam sites where early hopes for high status, cattle-richholdings seem to have been regularly frustrated by the realities of island

674 Lawson et al.

Fig. 10. The archaeofauna of Undir Junkarinsfløtti in comparison with approximatelycontemporary archaeofauna from Iceland and Greenland (indicative averages of data fromsites listed in the text).

farming. The large number of newborn (neonatal) cattle bones (20–50% ofall cattle) strongly suggests the same sort of dairy economy already doc-umented in Iceland, Greenland, and the Northern and Western Isles ofScotland (Bond, 2002; McGovern, 1985; McGovern et al., 2001).

With the exception of a single bone in UJF2, all identified caprines inthe Undir Junkarinsfløtti collection are sheep. In landnam-era Iceland andGreenland, goats were far more common. In Iceland goats only declined

Fig. 11. Changing proportions of domestic mammal bones from the 2003 Undir Junkarinsfløttiarchaeofauna. Small percentages of goat and dog bones in two samples are arrowed.


to their modern ‘trace’ levels in the early thirteenth century A.D., and inGreenland goats remained nearly as common as sheep in many collectionsdown to the end of the colony in the fourteenth–fifteenth century A.D. Asgoats are more effective at metabolizing twigs and leaves than sheep, theirearly reduction in the Faroes may be tied to the absence of significant wood-land. Reconstructed age distributions suggest that sheep were managed formilk as much as for wool.

Pigs make up a substantial part of the Undir Junkarinsfløtti archae-ofauna throughout the sequence. The presence of substantial numbers ofpigs is commonplace in landnam sites in Iceland and Greenland (McGovernet al., 2001), but pigs rarely survived as a major element in the domesticeconomy much beyond the mid-eleventh century A.D. It is not knownexactly when pigs become extinct in the Faroes, but there is no mentionof pigs or pig-breeding in the special Faroese enactment Seyabrævi (“Sheepletter”) of A.D. 1298, which describes only the traditional, historically knownsheep-breeding of recent times (Brandt, 1996). Economic pig-keeping re-quires either woodland or marshland for pannage, neither of which waspresent in the Faroes, or a source of fodder for penned, sty-kept animals(Ward and Mainland, 1999). Place-name evidence (see below) supports thenotion that pigs were closely herded and controlled. Fish offal has been sug-gested as a readily available source of fodder for pigs (McGovern, 1985),but stable carbon isotopic measurements on pig bones from both UJF1 andUJF2 indicate a wholly terrestrial diet. The role of pig keeping in the domes-tic economy of the ninth–thirteenth century Sandoy remains unclear, butthe persistence of pig keeping indicates an overall farming strategy differentfrom either contemporary Iceland and Greenland or the sheep-dominatedherding of recent centuries.

LANDSCAPE ARCHAEOLOGY ANDPLACE-NAME EVIDENCE

Sandoy is littered with archaeological sites and place-names that canprovide insights into the human use, impact on, and adaptation of the widerlandscape. These archaeological features are concentrated within the in-fields but are usually post-Medieval and modern in age, obscuring the ear-lier remains. By contrast, the outfield areas preserve much older features(Arge, in press). Here, the area around Lıtlavatn (Fig. 1) will be taken as anexample to demonstrate the layers of inference possible.

Today, the area mainly represents the outfield of the villages ofSkalavık and Husavık, and land use is characterized by sheep farming andpeat extraction. Post-Medieval and modern structures relating to sheepfarming occur all over the landscape along with extensive peat cuttings.

676 Lawson et al.

More ephemeral turf features, such as a series of low degraded walls westof Lıtlavatn, probably relate to animal husbandry in earlier periods. To thenorth of the lake is a small ruin complex and associated enclosure calledElinarhus (Elin’s House). The name of the site is traditionally associatedwith Elin, a servant of Servin, a mid-seventeenth century farmer in Husavık.Elin is said to have stayed here when she was looking after the grazing cattlein the summer months. The upstanding remains represent the final phase inthe site’s history, as degraded and robbed wall-lines of earlier structures oc-cur below the latest phase of construction. Therefore, the site is multiperiodand could have served multiple uses over the centuries.

Throughout the area are place-names associated with animal hus-bandry (e.g., Okasagil relates to the herding of cattle). In several areas,the place-name -byrgi is found, sometimes associated with a structure. Abyrgi is an area enclosed either by topography or artificial structures, whichmay have been used for different purposes, including collecting animals.The -byrgi place names are particularly concentrated on the relatively flatground to the west of Lıtlavatn. Another place name related to farmingpractices is Hoygarshellurnar (“the paving of the hay-enclosure”). Oral tra-dition suggests that most of the low turf walls and enclosures relate tosheep-farming, but place names are also present indicating ancient pig-farming (e.g., Svınsstøuheyggjurin [“the mound by the site where pigs arecollected”] and Svınstıaheyggjarnir [“the mounds by the path along whichthe pigs are driven”]). Both these place-names relate to areas with no ar-chaeological remains, but in the outfield of the village of Skarvanes, approx-imately 1.5 km southwest of Lıtlavatn, lies Stıggjurin a Svınhusinum, whichindicates a building, no longer extant, connected to pig-farming. It seemslikely, therefore, that many of the place-names have a long antiquity andrelate to economic strategies in place during the Norse and early medievalperiods when pig keeping was still part of the farming economy.

The movement and gathering of people through the landscape is alsoindicated by a number of place-names. For example, in the area aboveElinarhus to the north of the lake lies Byrgisgøta (“the path of the byrgi”).This sheep-path is always dry, even following heavy snow, and it would havebeen an important routeway in the past. Sornur, located on the west coast ofthe island where the waters from Storavatn leave the lake, indicates wherethe inhabitants from the island of Skuvoy came ashore to cut peat from theLıtlavatn area. Oral tradition suggests that the area near the river Tıfaldarabetween Lıtlavatn and Storavatn known as Millum Vatna (“between thelakes”) was used for the thing assembly for the island. The site lies approx-imately at the center of the island and at the border between the lands ofSkalavık, Husavık, and Sandur. Ancient outdoor thing assemblies are tra-ditionally thought to have occurred at many other sites elsewhere in the


Faroes. However, there is no archaeological or literary record to supportthis tradition.

DISCUSSION

The data reported here lead us to some initial conclusions concern-ing the impact of people on the environments of Sandoy. The landscapewe see today is much more similar to the natural, pre-settlement landscapethan in most other areas around the North Atlantic. Already prior to thesettlement there was no woodland, with at most a few scattered stands oftrees. Blanket peat had spread across large areas of the landscape, driven bypurely natural processes, with grass-sedge communities dominating on min-eral soils. Erosional processes were already very important in the landscape.

When the first settlers arrived, many components of the landscape sys-tem continued to operate in much the same way as before. There were noforests to remove; tall herbs and juniper were heavily reduced, but thisseems to have had little effect on the remaining vegetation. Blanket bogwas not encouraged by human activity, as elsewhere in the region—on thecontrary, grazing slightly reduced the dominance of ericaceous shrubs infavour of grasses, and over time much peat has been removed for use asfuel or fertilizer, the resulting peat cuttings being obvious in the landscapetoday. Infield vegetation and soils were modified by cultivation for crops,including at least the Hordeum registered in the Grothusvatn pollen recordand plant macrofossils from Undir Junkarinsfløtti, and also, we assume, forhay. Micro-charcoal is not abundant in our records, suggesting that burningwas not a major factor in landscape management—unlike for earlier pe-riods in Norway (Kaland, 1986) and the Western and Northern Isles ofScotland (Edwards, 1996), or in Norse Greenland (McGovern and Jordan,1982). Human impacts seem to have been concentrated in and around thevillage areas, in contrast to the more spatially extensive impacts of Vikingage settlements in Iceland and Greenland.

So it appears that the Sandoy landscape was modified by human impactto a surprisingly small degree in view of some attributes that might lead usto expect otherwise; namely, steep slopes, high rainfall, freeze-thaw activ-ity in the uplands, low vegetational productivity, and an environment thathad evolved to a dynamic equilibrium in the absence of grazing herbivores.There are at least four possible reasons for this ‘tough’ environmental re-sponse to the settlement.

The first is that the natural vegetation was insensitive to human impact.The major elements of the pre-settlement vegetation—grasses, sedges, andericaceous shrubs—are all capable of tolerating grazing, and indeed oftenprofited from the decline in woodland when herbivores were introduced to

678 Lawson et al.

other North Atlantic islands (e.g., Bennett et al., 1992; Hallsdottir, 1987).Only the tall herbs, and small populations of juniper and tree birch, seemto have suffered from grazing. The significance of this is that the vegeta-tion was not structurally altered. An illustrative contrast is Iceland, wherebirch woodland was the typical lowland vegetation prior to the arrival ofhumans. Some combination of grazing, land clearance, and exploitation forfuel and timber quickly led to widespread deforestation (e.g., Hallsdottir,1987). The removal of the dominant species in the ecosystem had substan-tial secondary effects on many aspects of the environment: its hydrologyand pedology were altered, with consequences for soil erosion and trans-port (Simpson et al., 2001); understorey fauna and flora were replaced byopen heath and grass/sedge communities (e.g., Hallsdottir, 1987). From thepoint of view of the settlers, their resource base was fundamentally altered,and the whole character of cultural landscape perception, routes of move-ment, available settlement areas, and viewshed were changed forever bythe creation of an open, largely unwooded landscape. As Vesteinsson et al.(2002) argue, many Icelandic valley bottoms became attractive for settle-ment only after the clearance of the woodlands. In the Faroes, the domi-nant taxa in the pre-landnam ecosystem happened to be resistant to humanimpact; a brittle collapse and fundamental reorganization of the ecosystem,as seen in Iceland, was thereby avoided.

Secondly, intensive land use was limited in spatial extent by the topog-raphy and climatic conditions of the island. Cultivation did make profoundchanges to the vegetation and soils (Adderley and Simpson, 2005), but onlyin the well drained, gently to moderately sloping (Arge, 1991), fertile areasaround settlements. At higher elevations, on very steep slopes, and on wa-terlogged soils and peat, cultivation was not a practical option, so impactswere limited both in kind and in degree.

Thirdly, Sandoy’s soil system was relatively robust in the face of hu-man impact. The extreme opposite case is that of Iceland, where andisolsbuilt up during the Holocene mainly by the accumulation of fine volcanicsediments, held together by a fragile vegetation cover. Once the vegetationwas damaged or removed and the soils exposed to the wind, erosion oc-curred catastrophically in many areas (e.g., Olafsdottir, 2001). By contrast,Faroese soils, including those of Sandoy, are dominated by silt- and sand-sized silicates that are generally more resistant to transportation and ero-sion. Those areas of the landscape that were susceptible to erosion, such asthe Lıtlavatn catchment, had already begun to erode in the mid-Holocene.Consequently, the first settlers encountered an environment which was in astate of dynamic equilibrium, energetically stable, without the potential fora vast threshold-driven collapse of a metastable soil system, as in Iceland.


While the three preceding explanations rest on inherent properties ofthe landscape, the fourth concerns the way in which humans managed thelandscape. Human settlement and economy on Sandoy, and perhaps in theFaroes generally, seem to have taken a very different path from those ofIceland and Greenland. Barley and hay cultivation were comparatively suc-cessful on Sandoy, thanks to the benign climate and fertile soils (madeall the more fertile by careful manuring and/or crop rotation; Adderleyand Simpson, 2005). In Iceland and Greenland, barley production wasfar more difficult, such that by the mid-thirteenth century the NorwegianKing’s Mirror could state that most Greenlanders “had never seen bread”(Larson, 1917). Consequently, domestic stock took on an all-important rolein providing for household provisioning and meeting rent, tithe, or trib-ute obligations, and ninth–thirteenth century Icelanders and Greenlanderswere increasingly forced to rely on land-extensive pastoralism, often in-volving unsupervised grazing by substantial flocks of sheep and goats(Perdikaris and McGovern, in press). Meanwhile, the medieval Faroese en-joyed improving yields from their increasingly fertile, intensively cultivatedbarley fields, set in an increasingly structured and controlled social land-scape. If Faroese farmers did choose to limit the size of their domestic flocksand herds, managing them closely in the context of an evolving partir sys-tem, and retaining pigs as part of a more diverse economy, they may haveneeded to provision themselves with a sustained (and clearly sustainable)harvest of sea birds, supplemented by marine fish. The available palaeoe-conomic evidence strongly indicates the early adoption of a distinctivelyFaroese approach to economy and land management which had profoundimplications for the subsequent evolution of both environment and society.

CONCLUSIONS

1. Many key elements of the modern Sandoy landscape—the absenceof trees, the large areas of blanket mire, and high rates of erosionin the uplands and on slopes—were already well established by thetime the first settlers arrived on the island.

2. A number of changes occurred in the environment following thesettlement, including a decline in certain plant species (Juniperus,Calluna) and an expansion of others (Poaceae, Plantago), probablythrough a combination of grazing and cultivation; active processesof erosion were accelerated; peat was extracted for fuel and otherwild resources, especially birds, were exploited; and structures andpaths were constructed.

680 Lawson et al.

3. By comparison with other North Atlantic situations, especiallyIceland, Greenland, and Atlantic Scotland, the change in the land-scape brought about by settlement was slight.

4. In many respects, the pre-settlement landscape already resembledthat of an anthropogenically “perturbed” North Atlantic island.Some potential human impacts, such as deforestation, were impos-sible, and others, such as those leading to enhanced soil erosion,merely amplified extant processes.

5. Archaeological data suggest that the economy of the colonists andtheir descendants departed substantially from the trajectory fol-lowed in contemporary Iceland and Greenland. Perhaps sustainedby an increasingly successful barley crop, the early Faroese appearto have de-emphasized animal husbandry and sustainably exploitedsea bird colonies for centuries, rather than rapidly drawing downand expending their natural capital. While there is still much tolearn about the interacting palaeoeconomy and palaeoecology ofthe Faroe Islands, it is clear that the historical ecology of this Viking-age settlement, isolated by the North Atlantic, provides an impor-tant and exceptional case of successful sustained management ofland and resources.

ACKNOWLEDGMENTS

The authors thank the Leverhulme Trust Landscapes circum-landnamprogramme, the US National Science Foundation Arctic Social SciencesProgram, and the CUNY Northern Science and Education Centre forfunding support. We thank Ragnar Edvardsson, Matt Brown, KateKrivogorskaya, Seth Brewington, Ramona Harrison, and Juha Martillafor their hard and expert work in the field in 2003–04. Thanks are alsodue to Julie Bond and Steve Dockrill for their good advice and council,to Paul Buckland for comments on the text, to Jill Barber, Iona Grant,Maureen Lamb, and Julie Mitchell for technical and logistical support, andHarald Jensen, the people of Sandur and Sandoy, and their mayor Palla Reynatugvu, for kind permission for access to sites and substantial helpwith excavations.

REFERENCES

Adderley, W. P., and Simpson, I. A. (2005). Early-Norse home-field productivity in the FaroeIslands. Human Ecology 33(5): 711–735.

Amorosi, T. (1992). Climate impact and human response at Svalbar farm, NE Iceland. InMorris, C. D., and Rackham, J. (eds.), Norse and Later Settlement and Subsistence in theNorth Atlantic, Glasgow University Press, Glasgow, pp. 131–152.


Amorosi, T., Woollett, J. W., Perdikaris, S., and McGovern, T. H. (1996). Regional zooar-chaeology and global change research: Problems and potentials. World Archaeology 28:126–157.

Arge, S. V. (1991). The Landnam in the Faroes. Arctic Anthropology 28: 101–120.Arge, S. V. (2001). Forn busetning heima a Sandi. Frø i 2: 5–13.Arge, S. V. (in press). Uttangar s, a characteristic of relics in the Faroe outfield. In Proceedings

of the Interdisciplinary Conference: “Utmark” (The Outfield/Wasteland)—as Industry andIdeology in the Iron Age and Middle Ages.” September 25–28, 2003, Bergen, Norway.

Barlow, L. K., Sadler, J. P., Ogilvie, A. E. J., Buckland, P. C., Amorosi, T., Ingimundarson,J. H., Skidmore, P., Dugmore, A. J., and McGovern, T. H. (1997). Interdisciplinary in-vestigations of the end of the Norse western settlement in Greenland. The Holocene 7:489–499.

Bennett, K. D., and Willis, K. J. (2002). Pollen. In Smol, J. P., Birks, H. J. B., and Last, W. M.(eds.), Tracking Environmental Change Using Lake Sediments, Vol. 3: Terrestrial, Algaland Siliceous Indicators, Kluwer, Dordrecht, pp. 5–32.

Bennett, K. D., Boreham, S., Sharp, M. J., and Switsur, V. R. (1992). Holocene history of envi-ronment, vegetation and human settlement on Catta Ness, Lunnasting, Shetland. Journalof Ecology 80: 241–273.

Bennett, K. D., Bunting, M. J., and Fossitt, J. A. (1997). Long-term vegetation change in theWestern and Northern Isles, Scotland. Botanical Journal of Scotland 49: 127–140.

Bond, J. (2002). Pictish pigs and Celtic cowboys: Food and farming in the Atlantic Iron Age.In Ballin Smith, B., and Banks, I. (eds.), In the Shadow of the Brochs: the Iron Age inScotland, Tempus Publishing, Stroud, pp. 177–184.

Brandt, J. (1996). Sheep breeding on Eastern Sandoy. In Guttesen, R. (ed.), The FaeroeIslands Topographic Atlas, Det Kongelige Danske Geografiska Selskab and Kort &Matrikelstyrelsen, Copenhagen, pp. 82–84.

Brayshay, B. A., Gilbertson, D. D., Kent, M., Edwards, K. J., Wathern, P., and Weaver, R. E.(2000). Surface pollen-vegetation relationships on the Atlantic seaboard: South Uist, Scot-land. Journal of Biogeography 27: 359–378.

Bronk Ramsey, C. (2003). OxCal Version 3.9. (http: //www.units.ox.ac.uk/departments/rlaha/orau/06 ind.html).

Brooks, S. J., Bennion, H., and Birks, H. J. B. (2001). Tracing lake trophic history with achironomid-total phosphorus inference model. Freshwater Biology 46: 513–533.

Buckland, P. C., Edwards, K. J., Sadler, J. P., and Dinnin, M. H. (1998). Late Holocene insectfaunas from Mykines, Faroe Islands, with observations on associated pollen and earlysettlement records. Fro skaparrit 48: 287–296.

Bunting, M. J. (1996). The development of heathland in Orkney, Scotland: Pollen recordsfrom Loch of Knitchen (Roussay) and Loch of Torness (Hoy). The Holocene 6: 193–212.

Bunting, M. J. (2003). Pollen-vegetation relationships in non-arboreal moorland taxa. Reviewof Palaeobotany and Palynology 125: 285–298.

Charman, D. J. (1992). Blanket mire formation at the Cross Lochs, Sutherland, northernScotland. Boreas 21: 53–72.

Charman, D. J. (2002). Peatlands and Environmental Change, Wiley, Chichester.Church, M. J. (2002). Plants and people in the later prehistoric and Norse periods of the Western

Isles of Scotland, Unpublished PhD Thesis, University of Edinburgh.Church, M. J., and Peters, C. (2004). Application of mineral magnetism in Atlantic Scotland

archaeology 2: Archaeobotanical taphonomy in Atlantic Scotland. In Housley, R., andColes, G. M. (eds.), Atlantic connections and adaptations: economies, environments andsubsistence in lands bordering the North Atlantic, Oxbow, Oxford, pp. 99–115.

Church, M. J., Arge, S. V., Brewington, S., McGovern, T. H., Woollett, J. W., Perdikaris, S.,Lawson, I. T., Amundsen, C., Harrison, R., and Krivogorskaya, K. (2005). Puffins, pigs,cod, and barley: Palaeoeconomy at Undir Junkarinsfløtti, Sandoy, Faroe Islands. Envi-ronmental Archaeology 10: 179–197.

Crawford, R. M. M. (2000). Ecological hazards of oceanic environments. New Phytologist 147:257–281.

682 Lawson et al.

Dean, W. E. (1974). Determination of carbonate and organic matter in calcareous sedimentsand sedimentary rocks by loss-on-ignition: Comparison with other methods. Journal ofSedimentary Petrology 44: 242–248.

Dearing, J. A. (1999). Magnetic susceptibility. In Walden, J., Oldfield, F., and Smith, J. P.(eds.), Environmental Magnetism: A Practical Guide, Technical Guide No. 6, QuaternaryResearch Association, London, pp. 35–62.

Dennis, A., Foote, P., and Perkins, R. (2000). The Laws of Early Iceland Gragas, Vol. II,University of Manitoba Press, Winnipeg.

Dickson, C. A. (1998). Past uses of turf in the Northern Isles. In Milles, C., and Coles,G. M. (eds.), On the Edge: Settlement in Marginal Areas, Oxbow, Oxford, pp. 105–109.

Dickson, J. H. (1992). North American driftwood, especially Picea (spruce), from archaeo-logical sites in the Hebrides and Northern Isles of Scotland. Review of Palaeobotany andPalynology 73: 49–56.

Dugmore, A. J., and Buckland, P. C. (1991). Tephrochronology and late Holocene soil ero-sion in Iceland. In Maizels, J., and Caseldine, C. (eds.), Environmental Change in Iceland,Kluwer, Dordrecht, pp. 147–159.

Dugmore, A. J., Buckland, P. C., Church, M. J., Dawson, A., Edwards, K. J., Keller, C.,Mayewski, P., McGovern, T. H., Mairs, K.-A., and Sveinbjarnardottir, G. (in press). High-resolution proxy climate records from the Greenland icesheet, settlement and landscapechange in the North Atlantic islands. Human Ecology 33(5).

Edwards, K. J. (1996). A Mesolithic of the Western and Northern Isles of Scotland? Evidencefrom pollen and charcoal. In Pollard, T., and Morrison, A. (eds.), The Early Prehistory ofScotland, Edinburgh University Press, Edinburgh, pp. 23–38.

Edwards, K. J., Borthwick, D., Cook, G., Dugmore, A. J., Mairs, K.-A., Church, M. J.,Simpson, I. A., and Adderley, W. P. (2005). A hypothesis-based approach to landscapechange in Su uroy, Faroe Islands. Human Ecology 33(5): 621–650.

Edwards, K. J., and Whittington, G. (2001). Lake sediments, erosion and landscape changeduring the Holocene in Britain and Ireland. Catena 42: 143–173.

Enghoff, I. B. (2003). Hunting, fishing, and animal husbandry at the Farm Beneath the Sand,Western Greenland: An archaeozoological analysis of a Norse farm in the Western Settle-ment, Meddelelser om Grønland Man and Society 28, Copenhagen.

Fosaa, A. M. (2000). Wildflowers in the Faroe Islands, Føroya Natturugripasavn, Torshavn.Fosaa, A. M. (2001). A review of plant communities of the Faroe Islands. Fro skaparrit 48:

41–54.Grime, J. P., Hodgson, J. G., and Hunt, R. (1988). Comparative Plant Ecology: A Functional

Approach to Common British Species, Unwin Hyman, London.Guttesen, R. (2001). Plant production of a Faeroese farm 1813–1892, related to climatic fluc-

tuations. Geografisk Tidsskrift, Danish Journal of Geography 101: 67–76.Hallsdottir, M. (1987). Pollen analytical studies of human influence on vegetation in rela-

tion to the landnam tephra layer in Southwest Iceland, LUNDQUA Thesis 18, LundUniversity.

Hannon, G. E., and Bradshaw, R. H. W. (2000). Impacts and timing of the first human settle-ment on vegetation of the Faroe islands. Quaternary Research 54: 404–413.

Hannon, G. E., Hermanns-Au ardottir, M., and Wastegard, S. (1998). Human impact atTjørnuvık on the Faroe Islands. Fro skaparrit 46: 215–228.

Hannon, G. E., Wastegard, S., Bradshaw, E., and Bradshaw, R. H. W. (2001). Human impactand landscape degradation on the Faroe Islands. Biology and Environment: Proceedingsof the Royal Irish Academy 101B: 129–139.

Humlum, O., and Christiansen, H. H. (1998). Late Holocene climatic forcing of geomorphicactivity in the Faroe Islands. Fro skaparrit 46: 169–189.

Johansen, J. (1971). A Palaeobotanical study indicating a pre-viking settlement in Tjørnuvık,Faroe Islands. Fro skaparrit 19: 147–157.

Johansen, J. (1975). Pollen diagrams from the Shetland and Faroe Islands. New Phytologist 75:369–387.


Johansen, J. (1982). Vegetational development in the Faroes from 10,000 BP to the present.Danmarks Geologiske Undersøgelse Arbog 1981: 111–136.

Johansen, J. (1985). Studies in the vegetational history of the Faroe and Shetland Islands, FøroyaFro skaparfelag, Torshavn.

Johansen, J. (1989). Survey of geology, climate and vegetation history. In Højgaard, A.,Johansen, J., and Odum, S. (eds.), A Century of Tree-planting in the Faroe Islands, FøroyaFro skaparfelag, Torshavn, pp. 11–15.

Jones, M. K. (1991). Sampling in palaeoethnobotany. In van Zeist, W., Wasylikowa, K.,and Behre, K. E. (eds.), Progress in Old World Palaeoethnobotany, A. A. Balkema,Rotterdam, pp. 53–62.

Kaland, P. E. (1986). The origin and management of Norwegian coastal heaths as reflected bypollen analysis. In Behre, K. E. (ed.), Anthropogenic Indicators in Pollen Diagrams, A.A.Balkema, Rotterdam, pp. 19–36.

Kristjanson, L. (1980). Islenzkir Sjavarhættir I, Bokautgafa Menningarsjo s, Reykjavık.Lang, B., Bedford, A. P., Richardson, N., and Brooks, S. J. (2003). The use of ultra-sound

in the preparation of carbonate and clay sediments for chironomid analysis. Journal ofPaleolimnology 30: 451–460.

Larson, L. M. (1917). The King’s Mirror, Twayne, New York.Malmros, C. (1990). Viking Age wood resources at Argisbrekka, Faroe Islands. Norwegian

Archaeological Review 23: 86–92.Malmros, C. (1994). Exploitation of local, drifted and imported wood by the Vikings on the

Faroe Islands. Botanical Journal of Scotland 46: 552–558.McGovern, T. H. (1985). Contributions to the Paleoeconomy of Norse Greenland. Acta

Archaeologica 54: 73–122.McGovern, T. H. (1997). Climate change and created vulnerability in Norse Greenland. Cli-

mate and man in the Arctic: Proceedings from a seminar held by the Danish Polar Center,27 March, 1996, Danish Polar Center, Copenhagen, pp. 34–43.

McGovern, T. H., Buckland, P. C., Sveinbjarnardottir, G., Savory, D., Skidmore, P., andAndreasen, C. (1983). A study of the faunal and floral remains from two Norse farmsin the Western Settlement, Greenland. Arctic Anthropology 20: 93–120.

McGovern, T. H., Bigelow, G. F., Amorosi, T., Woollett, J. W., and Perdikaris, S. (1993).Animal bones from E17a Narsaq. In Vebaek, C. L. (ed.), Narsaq: A Norse LandnamaFarm, Meddelelser om Grønland Man and Society 18, Copenhagen.

McGovern, T. H., Amorosi, T., Perdikaris, S., and Woollett, J. W. (1996). Zooarchaeology ofSandnes V51: Economic Change at a Chieftain’s Farm in West Greenland. Arctic Anthro-pology 33: 94–122.

McGovern, T. H., and Jordan, R. H. (1982). Settlement and landuse in the inner fjords ofGodthaab District, West Greenland. Arctic Anthropology 19: 63–80.

McGovern, T. H., Perdikaris, S., and Tinsley, C. (2001). Economy of Landnam: The evidenceof zooarchaeology. In Wawn, A., and Sigur ardottir, T. (eds.), Approaches to Vinland,Nordahl Institute Studies 4, Reykjavık, pp. 154–166.

Meeker, L. D., and Mayewski, P. A. (2002). A 1400 year long record of atmospheric circulationover the North Atlantic and Asia. The Holocene 12: 257–266.

Moore, P. D. (1975). Origin of blanket mires. Nature 256: 267–269.Moore, P. D. (1993). The origin of blanket mire, revisited. In Chambers, F. M. (ed.), Cli-

mate Change and Human Impact on the Landscape, Chapman and Hall, London, pp. 217–224.

Nolan, A. J., Henderson, D. J., and Merrell, B. G. (1995). The vegetation dynamics of wetheaths in relation to sheep grazing intensity. In Thompson, D. B. A., Hester, A. J., andUsher, M. B. (eds.), Heaths and Moorland: Cultural Landscapes, HMSO, Edinburgh,pp. 174–179.

Olafsdottir, R. (2001). Land degradation and climate in Iceland: a spatial and temporal assess-ment. PhD thesis, Lund University.

Outram, A. K. (1999). A Comparison of Palaeoeskimo and Medieval Norse Bone fat Exploita-tion in Western Greenland. Arctic Anthropology 36: 103–117.

684 Lawson et al.

Outram, A. K. (2003). Comparing levels of subsistence stress amongst Norse settlers in Ice-land and Greenland using levels of bone fat exploitation as an indicator. EnvironmentalArchaeology 8: 119–128.

Perdikaris, S., Amundsen, C., and McGovern, T. H. (2002). Report of Animal Bones fromTjarnargata 3C, Reykjavık, Iceland, unpublished report submitted to FornleifastofnanIslands, Reykjavık.

Perdikaris, S., and McGovern, T. H. (in press). Cod Fish, Walrus, and Chieftains: EconomicIntensification in the Norse North Atlantic. In Thurston, T. (ed.), New Perspectives onIntensification, Plenum, New York.

Randall, R. E., Andrew, R., and West, R. G. (1986). Pollen catchment in relation to localvegetation: Ceann Ear, Monach Isles N. N. R., Outer Hebrides. New Phytologist 104:271–310.

Rasmussen, J. (1982). The Faeroe Islands: Geology. In Rutherford, G. K. (ed.), The PhysicalEnvironment of the Faeroe Islands, Monographiae Biologicae Volume 46, Junk Publish-ers, The Hague, pp. 13–34.

Simpson, I. A., Dugmore, A. J., Thomson, A., and Vesteinsson, O. (2001). Crossing the thresh-olds: Human ecology and historical patterns of landscape degradation. Catena 42: 175–192.

Simpson, I. A., Gu mundsson, G., Thomson, A. M., and Cluett, J. (2004). Assessing the role ofwinter grazing in historic land degradation, Myvatnssveit, northeast Iceland. Geoarchae-ology 19: 471–502.

Small, A. (1992). The juniper decline during the Norse landnam in the Faroe Islands. ActaBorealia 91: 3–7.

Sobey, D. G. (1981). Biological flora of the British Isles. Stellaria media (L.) Vill. Journal ofEcology 69: 311–335.

Solem, T. (1989). Blanket mire formations at Haramsøy, Møre og Romsdal, Western Norway.Boreas 18: 221–235.

Stevenson, A. C., and Thompson, D. B. A. (1993). Long-term changes in the extent of heathermoorland in upland Britain and Ireland: Palaeoecological evidence for the importance ofgrazing. The Holocene 3: 70–76.

Stace, C. (1997). New flora of the British Isles, 2nd edn. Cambridge University Press,Cambridge, UK.

Stuiver, M., Reimer, P. J., Bard, E., Beck, J. W., Burr, G. S., Hughen, K. A., Kromer, B.,McCormac, G., Van der Plicht, J., and Spurk, M. (1998). INTCAL98 radiocarbon agecalibration, 24,000–0 cal BP. Radiocarbon 40: 1041–1083.

Tyldesley, J. B. (1973). Long-range transmission of tree pollen to Shetland. III. frequenciesover the past hundred years. New Phytologist 72: 691–698.

Vesteinsson, O., McGovern, T. H., and Keller, C. (2002). Enduring Impacts: Social and En-vironmental Aspects of Viking Age Settlement in Iceland and Greenland. ArchaeologiaIslandica 2: 98–136.

Vickers, K., Bending, J., Buckland, P. C., Edwards, K. J., Hansen, S. S., and Cook, G. (2005).Toftanes: The paleoecology of a Faroese Landnam farm. Human Ecology 33(5): 685–710.

Ward, J., and Mainland, I. (1999). Microwear in modern rooting and stall fed pigs. Environ-mental Archaeology 4: 25–33.

Wastegard, S. (2002). Early to middle Holocene silicic tephra horizons from the Katla volcanicsystem, Iceland: New results from the Faroe Islands. The Holocene 2002: 723–730.

Woollett, J. W., Arge, S. V., Church, M. J., and McGovern, T. H. (2004). A report of archaeo-logical fieldwork at Undir Junkarinsfløtti, Sandur, Sandoy, 2003. Unpublished Landscapescircum-Landnam Project Field Report.

historical ecology on sandoy, faroe islands ...€¦ · change on sandoy, faroe islands, especially...

Documents